AcademicMomentum: Machine Learning for Grade Prediction

AcademicMomentum in Action
Predicted Result of Student Data
Correlation Chart that the model generated

GITHUB REPOSITORY

https://github.com/liaiwen8/AcademicMomentum

YOUTUBE PITCH PRESENTATION

Inspiration

What inspired me to create AcademicMomentum was a deep-seated belief in the transformative power of education. Having witnessed the challenges that both students and educators face in our current educational systems, I became determined to find a way to make a meaningful difference. I wanted to create a tool that not only empowers students to reach their full potential but also give educators valuable insights.

What it does

AcademicMomentum is a powerful application that utilizes a machine learning model to accurately predict a student's final grade based on their past performance and environment. It takes variables as input and forecasts a student's future grade. It then visualizes correlations and data in graphs.

How we built it

The machine learning model uses a Multiple Random Tree Regression algorithm to regress a set of values and generate forecasted interpolated values. It does this by training itself on a dataset of the grades and outcomes of Portuguese university students, then predicting the results of new data. Accuracy ranges from 86% to 93%.

Challenges I ran into

I first implemented a linear regression model, which generated low accuracy. I learned how to predict regressed values with a new algorithm (Random Tree Regression), and this greatly improved the accuracy of my model. I did not know how to code a front-end GUI for my application, and I did not want users to struggle with a Command-Line Interface, so I had to learn and debug my own GUI in Python (I am fluent in C and base most of my projects in it, so this was a new experience)

Impact

AcademicMomentum will give students a dynamic tool for self-reflection and empower their mindsets for success. It will help them set viable and impactful study goals, and give insight into their academic performance. AcademicMomentum will also benefit teachers, by allowing them to easily identify which students need their help the most, as some students may be afraid to ask for help or do not know how to. Additionally, many students will have doors closed on their futures based on which high school pathways and course streaming their teachers pick. This may not be the fault of the student in question, and they may be very willing to work hard and persevere through challenges. AcademicMomentum and its intelligent machine learning model will be able to reassure educators that students who may seem weak have the power to change their lives. Most importantly, AcademicMomentum will increase education efficiency by giving teachers knowledge into how their class is understanding the material.

Future Research

I aim to partner with educational institutions to collect more personalized data to Calgary schools and train my model for students in middle or elementary school. Real world testing is extremely important for machine learning models. Additionally, I want to upgrade the application to provide customized recommendations for students and teachers on what actions to take to further their success, using the AI model. I also aim to give users (particularly students) personalized study plans to prepare for a test.